{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T21:45:26Z","timestamp":1782855926028,"version":"3.54.5"},"reference-count":59,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-017"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-012"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T00:00:00Z","timestamp":1790812800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-004"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["61974018"],"award-info":[{"award-number":["61974018"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Neurocomputing"],"published-print":{"date-parts":[[2026,10]]},"DOI":"10.1016\/j.neucom.2026.134285","type":"journal-article","created":{"date-parts":[[2026,6,17]],"date-time":"2026-06-17T16:29:41Z","timestamp":1781713781000},"page":"134285","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":0,"special_numbering":"C","title":["fastiRNN: Fast inference of recurrent neural networks at the post-training stage"],"prefix":"10.1016","volume":"698","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-4003-2758","authenticated-orcid":false,"given":"Hai","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-7129-0209","authenticated-orcid":false,"given":"Jiaxuan","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yu","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiming","family":"Li","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"10.1016\/j.neucom.2026.134285_bib0005","doi-asserted-by":"crossref","first-page":"533","DOI":"10.1038\/323533a0","article-title":"Learning representations by back-propagating errors","volume":"323","author":"Rumelhart","year":"1986","journal-title":"Nature"},{"key":"10.1016\/j.neucom.2026.134285_bib0010","series-title":"Deep Learning","author":"Goodfellow","year":"2016"},{"issue":"4","key":"10.1016\/j.neucom.2026.134285_bib0015","doi-asserted-by":"crossref","first-page":"485","DOI":"10.1109\/JPROC.2020.2976475","article-title":"Model compression and hardware acceleration for neural networks: a comprehensive survey","volume":"108","author":"Deng","year":"2020","journal-title":"Proc. of the IEEE"},{"key":"10.1016\/j.neucom.2026.134285_bib0020","series-title":"Proc. Int. Conf. on Learning Representations (ICLR)","article-title":"Exploring sparsity in recurrent neural networks","author":"Narang","year":"2017"},{"key":"10.1016\/j.neucom.2026.134285_bib0025","series-title":"Proc. Int. Conf. on Learning Representations (ICLR)","article-title":"Learning intrinsic sparse structures within long short-term memory","author":"Wen","year":"2018"},{"key":"10.1016\/j.neucom.2026.134285_bib0030","doi-asserted-by":"crossref","first-page":"134","DOI":"10.1016\/j.neunet.2019.11.018","article-title":"Structured pruning of recurrent neural networks through neuron selection","volume":"123","author":"Wen","year":"2020","journal-title":"Neural Netw."},{"issue":"100605","key":"10.1016\/j.neucom.2026.134285_bib0035","first-page":"1","article-title":"A survey on knowledge distillation: recent advancements","volume":"18","author":"Moslemi","year":"2024","journal-title":"Mach. Learn. Appl."},{"key":"10.1016\/j.neucom.2026.134285_bib0040","series-title":"Proc. Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP)","first-page":"7230","article-title":"Knowledge distillation for recurrent neural network language modeling with trust regularization","author":"Shi","year":"2019"},{"key":"10.1016\/j.neucom.2026.134285_bib0045","doi-asserted-by":"crossref","first-page":"1541","DOI":"10.1007\/s11760-021-02108-9","article-title":"Knowledge distillation-based performance transferring for LSTM-RNN model acceleration","volume":"16","author":"Ma","year":"2022","journal-title":"Signal Image Video Process."},{"key":"10.1016\/j.neucom.2026.134285_bib0050","series-title":"Proc. Interspeech","first-page":"338","article-title":"Long short-term memory recurrent neural network architectures for large scale acoustic modeling","author":"Sak","year":"2014"},{"key":"10.1016\/j.neucom.2026.134285_bib0055","series-title":"Proc. British Machine Vision Conf. (BMVC)","article-title":"Speeding up convolutional neural networks with low rank expansions","author":"Jaderberg","year":"2014"},{"key":"10.1016\/j.neucom.2026.134285_bib0060","series-title":"Proc. IEEE Int. Conf. on Computer Vision (ICCV)","article-title":"Fast R-CNN","author":"Girshick","year":"2015"},{"key":"10.1016\/j.neucom.2026.134285_bib0065","series-title":"Proc. Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP)","first-page":"5960","article-title":"Learning compact recurrent neural networks","author":"Lu","year":"2016"},{"issue":"10","key":"10.1016\/j.neucom.2026.134285_bib0070","doi-asserted-by":"crossref","first-page":"1943","DOI":"10.1109\/TPAMI.2015.2502579","article-title":"Accelerating very deep convolutional networks for classification and detection","volume":"38","author":"Zhang","year":"2016","journal-title":"IEEE Trans. on Pattern Analysis and Machine Intelligence"},{"issue":"4","key":"10.1016\/j.neucom.2026.134285_bib0075","doi-asserted-by":"crossref","first-page":"1096","DOI":"10.1137\/19M1296070","article-title":"Active subspace of neural networks: structural analysis and universal attacks","volume":"2","author":"Cui","year":"2020","journal-title":"SIAM J. Math. Data Sci."},{"issue":"11","key":"10.1016\/j.neucom.2026.134285_bib0080","doi-asserted-by":"crossref","first-page":"4169","DOI":"10.1007\/s10115-020-01487-8","article-title":"Fast LSTM by dynamic decomposition on cloud and distributed systems","volume":"62","author":"You","year":"2020","journal-title":"Knowl. Inf. Syst."},{"key":"10.1016\/j.neucom.2026.134285_bib0085","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"29321","article-title":"DRONE: data-aware low-rank compression for large NLP models","author":"Chen","year":"2021"},{"issue":"12","key":"10.1016\/j.neucom.2026.134285_bib0090","doi-asserted-by":"crossref","first-page":"10487","DOI":"10.1109\/TNNLS.2022.3167466","article-title":"fastESN: fast echo state network","volume":"34","author":"Wang","year":"2023","journal-title":"IEEE Trans. on Neural Networks and Learning Systems"},{"issue":"1","key":"10.1016\/j.neucom.2026.134285_bib0095","doi-asserted-by":"crossref","first-page":"519","DOI":"10.1109\/TNNLS.2022.3175757","article-title":"Accelerating neural ODEs using model order reduction","volume":"35","author":"Lehtim\u00e4ki","year":"2024","journal-title":"IEEE Trans. on Neural Networks and Learning Systems"},{"issue":"19","key":"10.1016\/j.neucom.2026.134285_bib0100","doi-asserted-by":"crossref","first-page":"22818","DOI":"10.1007\/s10489-023-04730-1","article-title":"A dimensionality reduction approach for convolutional neural networks","volume":"53","author":"Meneghetti","year":"2023","journal-title":"Appl. Intell."},{"key":"10.1016\/j.neucom.2026.134285_bib0105","series-title":"Nonlinear model reduction via discrete empirical interpolation","author":"Chaturantabut","year":"2011"},{"key":"10.1016\/j.neucom.2026.134285_bib0110","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"1135","article-title":"Learning both weights and connections for efficient neural network","author":"Han","year":"2015"},{"key":"10.1016\/j.neucom.2026.134285_bib0115","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"2082","article-title":"Learning structured sparsity in deep neural networks","author":"Wen","year":"2016"},{"key":"10.1016\/j.neucom.2026.134285_bib0120","series-title":"Proc. Int. Conf. on Learning Representations (ICLR)","article-title":"Pruning filters for efficient ConvNets","author":"Li","year":"2017"},{"issue":"10","key":"10.1016\/j.neucom.2026.134285_bib0125","doi-asserted-by":"crossref","first-page":"2525","DOI":"10.1109\/TPAMI.2018.2858232","article-title":"ThiNet: pruning CNN filters for a thinner net","volume":"41","author":"Luo","year":"2019","journal-title":"IEEE Trans. on Pattern Analysis and Machine Intelligence"},{"key":"10.1016\/j.neucom.2026.134285_bib0130","series-title":"Proc. IEEE\/CVF Conf. on Computer Vision and Pattern Recognition (CVPR)","first-page":"13","article-title":"Exploring the granularity of sparsity in convolutional neural networks","author":"Mao","year":"2017"},{"key":"10.1016\/j.neucom.2026.134285_bib0135","author":"Hinton"},{"issue":"4","key":"10.1016\/j.neucom.2026.134285_bib0140","doi-asserted-by":"crossref","DOI":"10.1137\/130916138","article-title":"Active subspace methods in theory and practice: applications to kriging surfaces","volume":"32","author":"Constantine","year":"2014","journal-title":"SIAM J. Sci. Comput."},{"key":"10.1016\/j.neucom.2026.134285_bib0145","doi-asserted-by":"crossref","first-page":"539","DOI":"10.1146\/annurev.fl.25.010193.002543","article-title":"The proper orthogonal decomposition in the analysis of turbulent flows","volume":"25","author":"Berkooz","year":"1993","journal-title":"Annu. Rev. Fluid Mech."},{"key":"10.1016\/j.neucom.2026.134285_bib0150","series-title":"Proc. Int. Conf. on Machine Learning (ICML)","first-page":"3891","article-title":"Tensor-train recurrent neural networks for video classification","author":"Yang","year":"2017"},{"key":"10.1016\/j.neucom.2026.134285_bib0155","series-title":"Proc. IEEE\/CVF Conf. on Computer Vision and Pattern Recognition (CVPR)","first-page":"9378","article-title":"Learning compact recurrent neural networks with block-term tensor decomposition","author":"Ye","year":"2018"},{"key":"10.1016\/j.neucom.2026.134285_bib0160","series-title":"Proc. of the Joint Int. Conf. on Computational Linguistics, Language Resources and Evaluation (LREC-COLING)","first-page":"10822","article-title":"Low-rank prune-and-factorize for language model compression","author":"Ren","year":"2024"},{"key":"10.1016\/j.neucom.2026.134285_bib0165","series-title":"Proc. of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD)","first-page":"6169","article-title":"Unified low-rank compression framework for click-through rate prediction","author":"Yu","year":"2024"},{"key":"10.1016\/j.neucom.2026.134285_bib0170","series-title":"Proc. Int. Conf. on Learning Representations (ICLR)","article-title":"Language model compression with weighted low-rank factorization","author":"Hsu","year":"2022"},{"issue":"12","key":"10.1016\/j.neucom.2026.134285_bib0175","doi-asserted-by":"crossref","first-page":"14403","DOI":"10.1007\/s10462-023-10496-2","article-title":"nmODE: neural memory ordinary differential equation","volume":"56","author":"Yi","year":"2023","journal-title":"Artif. Intell. Rev."},{"key":"10.1016\/j.neucom.2026.134285_bib0180","series-title":"Proc. Int. Joint Conf. on Artificial Intelligence (IJCAI)","first-page":"821","article-title":"A lightweight U-like network utilizing neural memory ordinary differential equations for slimming the decoder","author":"He","year":"2024"},{"key":"10.1016\/j.neucom.2026.134285_bib0185","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","article-title":"MobileODE: an extra lightweight network","author":"Yu","year":"2025"},{"key":"10.1016\/j.neucom.2026.134285_bib0190","series-title":"Proc. of Mathematical and Scientific Machine Learning Conf. (MSML), Vol. 107 of Proceedings of Machine Learning Research","first-page":"476","article-title":"Gating creates slow modes and controls phase-space complexity in GRUs and LSTMs","author":"Can","year":"2020"},{"issue":"8","key":"10.1016\/j.neucom.2026.134285_bib0195","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput."},{"key":"10.1016\/j.neucom.2026.134285_bib0200","series-title":"Proc. Workshop on Syntax, Semantics and Structure in Statistical Translation","first-page":"103","article-title":"On the properties of neural machine translation: encoder\u2013decoder approaches","author":"Cho","year":"2014"},{"key":"10.1016\/j.neucom.2026.134285_bib0205","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"493","article-title":"Hierarchical recurrent neural networks for long-term dependencies","author":"El Hihi","year":"1995"},{"key":"10.1016\/j.neucom.2026.134285_bib0210","series-title":"Proc. Int. Conf. on Learning Representations (ICLR)","article-title":"How to construct deep recurrent neural networks","author":"Pascanu","year":"2014"},{"issue":"4","key":"10.1016\/j.neucom.2026.134285_bib0215","doi-asserted-by":"crossref","first-page":"483","DOI":"10.1137\/130932715","article-title":"A survey of projection-based model reduction methods for parametric dynamical systems","volume":"57","author":"Benner","year":"2015","journal-title":"SIAM Rev."},{"issue":"3","key":"10.1016\/j.neucom.2026.134285_bib0220","doi-asserted-by":"crossref","first-page":"626","DOI":"10.1162\/NECO_a_00409","article-title":"Opening the black box: low-dimensional dynamics in high-dimensional recurrent neural networks","volume":"25","author":"Sussillo","year":"2013","journal-title":"Neural Comput."},{"key":"10.1016\/j.neucom.2026.134285_bib0225","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"435","article-title":"Preventing gradient explosions in gated recurrent units","author":"Kanai","year":"2017"},{"key":"10.1016\/j.neucom.2026.134285_bib0230","series-title":"Matrix Theory: Basic Results and Techniques","author":"Zhang","year":"2013"},{"key":"10.1016\/j.neucom.2026.134285_bib0235","series-title":"Proc. Design Automation Conf. (DAC)","article-title":"A stabilized discrete empirical interpolation method for model reduction of electrical, thermal, and microelectromechanical systems","author":"Hochman","year":"2011"},{"key":"10.1016\/j.neucom.2026.134285_bib0240","series-title":"The \u201cecho state\u201d approach to analysing and training recurrent neural networks","author":"Jaeger","year":"2001"},{"issue":"2","key":"10.1016\/j.neucom.2026.134285_bib0245","first-page":"313","article-title":"Building a large annotated corpus of English: the penn treebank","volume":"19","author":"Marcus","year":"1993","journal-title":"Comput. Linguist."},{"issue":"5","key":"10.1016\/j.neucom.2026.134285_bib0250","doi-asserted-by":"crossref","first-page":"2737","DOI":"10.1137\/090766498","article-title":"Nonlinear model reduction via discrete empirical interpolation","volume":"32","author":"Chaturantabut","year":"2010","journal-title":"SIAM J. Sci. Comput."},{"issue":"4","key":"10.1016\/j.neucom.2026.134285_bib0255","doi-asserted-by":"crossref","first-page":"1740","DOI":"10.1109\/TNNLS.2020.3043752","article-title":"Subtraction gates: another way to learn long-term dependencies in recurrent neural networks","volume":"33","author":"He","year":"2022","journal-title":"IEEE Trans. on Neural Networks and Learning Systems"},{"key":"10.1016\/j.neucom.2026.134285_bib0260","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"107547","article-title":"xLSTM: extended long short-term memory","author":"Beck","year":"2024"},{"key":"10.1016\/j.neucom.2026.134285_bib0265","author":"Paganini"},{"issue":"12","key":"10.1016\/j.neucom.2026.134285_bib0270","doi-asserted-by":"crossref","first-page":"10558","DOI":"10.1109\/TPAMI.2024.3447085","article-title":"A survey on deep neural network pruning: taxonomy, comparison, analysis, and recommendations","volume":"46","author":"Cheng","year":"2024","journal-title":"IEEE Trans. on Pattern Analysis and Machine Intelligence"},{"key":"10.1016\/j.neucom.2026.134285_bib0275","series-title":"Proc. Conf. on Neural Information Processing Systems (NeurIPS)","first-page":"301","article-title":"Recurrent networks and NARMA modeling","author":"Connor","year":"1991"},{"key":"10.1016\/j.neucom.2026.134285_bib0280","unstructured":"Jena climate dataset, www.bgc-jena.mpg.de\/wetter"},{"issue":"101819","key":"10.1016\/j.neucom.2026.134285_bib0285","first-page":"1","article-title":"Long sequence time-series forecasting with deep learning: a survey","volume":"97","author":"Chen","year":"2023","journal-title":"Inf. Fusion"},{"key":"10.1016\/j.neucom.2026.134285_bib0290","series-title":"Proc. AAAI Conf. on Artificial Intelligence (AAAI)","first-page":"11106","article-title":"Informer: beyond efficient transformer for long sequence time-series forecasting","author":"Zhou","year":"2021"},{"key":"10.1016\/j.neucom.2026.134285_bib0295","author":"Zaremba"}],"container-title":["Neurocomputing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0925231226016838?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0925231226016838?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T20:37:44Z","timestamp":1782851864000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S0925231226016838"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10]]},"references-count":59,"alternative-id":["S0925231226016838"],"URL":"https:\/\/doi.org\/10.1016\/j.neucom.2026.134285","relation":{},"ISSN":["0925-2312"],"issn-type":[{"value":"0925-2312","type":"print"}],"subject":[],"published":{"date-parts":[[2026,10]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"fastiRNN: Fast inference of recurrent neural networks at the post-training stage","name":"articletitle","label":"Article Title"},{"value":"Neurocomputing","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.neucom.2026.134285","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2026 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.","name":"copyright","label":"Copyright"}],"article-number":"134285"}}